There are, in fact, three, logically-distinct senses of the term “breakthrough infection”:
- Type 1: Vaccine-Escape Breakthrough. This is the type that everyone is—justifiably—scared of. A variant such as B.1.351, also now known as Beta emerges with a mutation that allows it to side-step vaccine-primed immunity to SARS-CoV-2. We were “fortunate” with Beta that it reproduced relatively slowly, so that it was outcompeted by Delta, which is in fact not a vaccine-escape variant, but which reproduces much more rapidly than Beta. [Note added 29 Dec 2021: As discussed in a later post in this series, I no longer believe that Beta was, in fact, a true vaccine-escape variant. In fact, I believe no true SARS-CoV-2 vaccine-escape variant has yet appeared.]
- Type 2: “Statistical” Breakthrough. Vaccines do not offer perfect protection. Even the near-miraculous mRNA vaccines such as Pfizer/BioNTech or Moderna have efficacies that are “only” about 95%. This means that on average, each infective contact with a vaccinated individual has a probability of producing an infection that is only 5% of what it would be for an unvaccinated individual. The probability is not zero. An efficacy number for a vaccine is, in effect, a prediction of a rate of “breakthrough infections”. This is one reason that “breakthrough infection” is such a wretched term: it hides an unspoken assumption that the vaccines constitute a perfect armor, and if a vaccinated individual becomes infected the virus must somehow have “broken through” the armor, which is idiotic. The real epidemiological question concerning dangerous variants producing Type 1 breakthrough infections is, always, “how many more vaccinated individuals are becoming infected than the rate predicted by the vaccine efficacy?” Somehow, this point is never communicated effectively.
- Type 3: False Breakthrough. What happens when a properly-functioning vaccine-primed immune system notices an infection and does its job to spec? The offending virus doesn’t just explode on contact. It enters the host and begins infecting cells and multiplying. Antibodies slow down its growth, and at some point the cellular immune system goes on alert, as T-cells come streaming out to clear the infection. The immune reaction itself may cause symptoms to appear, and an RT-PCR test for viral RNA, being an extremely sensitive assay, may in fact pick up evidence of the infection. Which is to say, such a vaccinated individual may be classified as a symptomatic, positive breakthrough case despite the fact that everything is going according to plan, and the individual will never experience more than mild flu-like symptoms, analogous to the reactions that some individuals experience to the vaccination shots themselves.
One might articulate the following objection to the Type 3 classification: a person who tests positive because their immune system has not yet cleared the infection might actually be infective, and this creates a risk to others even if they are not at great risk of disease themselves. The extent of this risk is a matter of debate (at least it was, until Omicron changed the situation). Part of the problem is that RT-PCR tests in fact measure concentration of viral RNA (by a measure called “Ct”), but almost never in fact release that value with a positive test result, so that we never know whether “positive” means “positive with a lot of virus” or “positive with trace amounts of virus”. I will be writing a separate post about why we don’t have that data, and the very considerable costs to epidemic surveillance of the resulting knowledge gap. My view is that through the advent of Delta (which took over the SARS-CoV-2 genome worldwide between January and August 2021) it was pretty clear from large vaccine effectiveness studies such as ours that there was not much evidence for “extra” breakthrough infections above and beyond the expected rates, so it looked as if mostly the vaccines were on the job.
One needs to be a little cautious, however. The trick that Delta learned is fast reproduction. Delta took over the genome by making many more copies of itself much faster than the previously prevalent strains during in-host infections. This has two effects: it shortens incubation periods, and makes viral loads (concentration of virus particles) in aerosols higher. Both of these effects accelerate transmission. Faster in-host reproduction also means that the virus could, in principle, reach infective levels of viral load in a Type 3 scenario before the cellular immune system puts on the clamps and wipes it out. In this case, such an infected individual would not be at much personal risk, but might expose others to infection risk. So it’s not ridiculous to imagine that a really fast-moving variant might require a reconsideration, or at least a re-nuancing, of the Type 3 classification.
Well, guess what? That really fast-moving variant is here. Omicron reproduces 5-6 times more rapidly than Delta, judging from the curves that I’ve seen. That means that even in a properly-functioning, vaccine-primed immune system that is perfectly capable of clobbering it, Omicron could in principle get up to infective viral load levels before getting wiped out. So, it seems to me, at a minimum, one needs a sub-classification of Type 3:
- Type 3a: “Classic” False Breakthrough. Nothing to see here. Just like the old story, a chimera produced by super-sensitive RT-PCR testing combined with unavailable data about viral load, in a situation where what really happened is that the vaccine-primed immune system did its job properly and the virus got clobbered before it could reach infective load levels;
- Type 3b: “Infective”, or Semi-Breakthrough. The virus was not sufficiently slowed down by the antibody defense layer of the adaptive immune system to prevent an infectious state from arising, but it is doomed nonetheless because the cellular immunity is enraged and on its way to wipe it out, which it will do in due course, well before any severe disease may develop.
So in the case of Omicron “breakthroughs” it is still important to isolate and prevent further contagion, of course. On the other hand, it is possible to see reasons why, despite all the “breakthrough” cases that are being attributed to this variant, people are not, in fact, dying like 1918 pandemic victims. Most Omicron breakthrough infections are, in all likelihood, either false breakthroughs or semi-breakthroughs. The key thing to understand is this: There is no evidence whatsoever that Omicron produces Type 1 breakthrough infections. None. In fact, there are rather good reasons to believe that the opposite is true. I’ll write a post soon about those reasons. For now, if I’ve persuaded you that the “breakthrough” story is complicated, and doesn’t always mean that there’s a wolf at the door, I think that’s a pretty good takeaway.